-----
edit as a reply to both responses:
Tests like this were mentioned in the article - but wouldn't the same effect be caused by people who had an honest belief that they saw the number 8 as orange, or if you sat someone in a room and played a tape for an hour that "One. Blue. Two. Violet. Three. Green.[...]" etc.?
i.e. if I've said that I see 5 as blue enough times, wouldn't that affect my times on an association test?
Are there any stronger examples of tests? Any papers that I should read? Honest question; not trying to be contrarian.